Artificial intelligence–assisted urine cytology based on the Paris System for Reporting Urothelial Carcinoma

IF 2.9 3区医学 Q3 ONCOLOGY

Cancer Cytopathology Pub Date : 2026-06-01 DOI:10.1002/cncy.70120

Lan Chen MD, PhD, Xinyi Cao BS, Zongyue Lu BS, Longteng Liu MSc, Mingjun Sun BS, Yulong Wu BS, Wei Zhang PhD

{"title":"Artificial intelligence–assisted urine cytology based on the Paris System for Reporting Urothelial Carcinoma","authors":"Lan Chen MD, PhD, Xinyi Cao BS, Zongyue Lu BS, Longteng Liu MSc, Mingjun Sun BS, Yulong Wu BS, Wei Zhang PhD","doi":"10.1002/cncy.70120","DOIUrl":null,"url":null,"abstract":"<div>\n \n \n <section>\n \n <h3> Background</h3>\n \n <p>Urine cytology is a noninvasive and valuable tool for detecting urothelial carcinoma but suffers from variable sensitivity and observer dependency. Artificial intelligence (AI) may enhance the diagnostic accuracy and efficiency of urine cytology. The objective of this study was to develop and validate an AI-based cytology system for urothelial carcinoma detection in both clinical and screening contexts.</p>\n </section>\n \n <section>\n \n <h3> Methods</h3>\n \n <p>In total, 328 retrospective clinical cases and 1489 prospective health screening samples were analyzed. All specimens underwent liquid-based cytology and were digitized at ×20 magnification. For model development, ed 269 annotated training slides (56,710 cells) were used. The AI pipeline mimicked cytopathologist workflow, integrating deep learning–based cell detection and segmentation with feature extraction and support vector machine classification into AI-negative, AI-atypical, and AI-positive categories according to The Paris System for Reporting the Urinary Cytology. Agreement was assessed using weighted κ values and prevalence-adjusted, bias-adjusted κ values.</p>\n </section>\n \n <section>\n \n <h3> Results</h3>\n \n <p>In the clinical cohort, AI achieved 83.0% agreement with histopathology and 82.6% agreement with expert cytology. Weighted κ values (κ, 0.631–0.661) reflected substantial agreement, with discrepancies mainly between adjacent categories. In the health screening cohort, the prevalence-adjusted, bias-adjusted κ was 0.647 despite low prevalence. AI demonstrated a high negative predictive value of 99.7% and negative percent agreement of 82.3%, with a minimal false-omission rate (0.25%).</p>\n </section>\n \n <section>\n \n <h3> Conclusions</h3>\n \n <p>AI-assisted urine cytology exhibited substantial concordance with expert interpretation and histopathologic standards. Its capability of screen-out and high negative predictive value support its potential as a reliable triage tool, improving workflow efficiency while maintaining diagnostic reliability in large-scale screening.</p>\n </section>\n </div>","PeriodicalId":9410,"journal":{"name":"Cancer Cytopathology","volume":"134 6","pages":""},"PeriodicalIF":2.9000,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Cancer Cytopathology","FirstCategoryId":"3","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/cncy.70120","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ONCOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

Background

Urine cytology is a noninvasive and valuable tool for detecting urothelial carcinoma but suffers from variable sensitivity and observer dependency. Artificial intelligence (AI) may enhance the diagnostic accuracy and efficiency of urine cytology. The objective of this study was to develop and validate an AI-based cytology system for urothelial carcinoma detection in both clinical and screening contexts.

Methods

In total, 328 retrospective clinical cases and 1489 prospective health screening samples were analyzed. All specimens underwent liquid-based cytology and were digitized at ×20 magnification. For model development, ed 269 annotated training slides (56,710 cells) were used. The AI pipeline mimicked cytopathologist workflow, integrating deep learning–based cell detection and segmentation with feature extraction and support vector machine classification into AI-negative, AI-atypical, and AI-positive categories according to The Paris System for Reporting the Urinary Cytology. Agreement was assessed using weighted κ values and prevalence-adjusted, bias-adjusted κ values.

Results

In the clinical cohort, AI achieved 83.0% agreement with histopathology and 82.6% agreement with expert cytology. Weighted κ values (κ, 0.631–0.661) reflected substantial agreement, with discrepancies mainly between adjacent categories. In the health screening cohort, the prevalence-adjusted, bias-adjusted κ was 0.647 despite low prevalence. AI demonstrated a high negative predictive value of 99.7% and negative percent agreement of 82.3%, with a minimal false-omission rate (0.25%).

Conclusions

AI-assisted urine cytology exhibited substantial concordance with expert interpretation and histopathologic standards. Its capability of screen-out and high negative predictive value support its potential as a reliable triage tool, improving workflow efficiency while maintaining diagnostic reliability in large-scale screening.

查看原文本刊更多论文

基于Paris系统报告尿路上皮癌的人工智能辅助尿细胞学。

背景：尿细胞学是检测尿路上皮癌的一种非侵入性和有价值的工具，但存在敏感性不稳定和观察者依赖性。人工智能（AI）可以提高尿细胞学诊断的准确性和效率。本研究的目的是开发和验证一种基于人工智能的细胞学系统，用于临床和筛查尿路上皮癌的检测。方法：对328例回顾性临床病例和1489例前瞻性健康筛查样本进行分析。所有标本都进行了液体细胞学检查，并在×20放大镜下进行了数字化处理。用于模型开发，使用了269张带注释的训练幻灯片（56,710个细胞）。人工智能管道模拟了细胞病理学家的工作流程，将基于深度学习的细胞检测和分割与特征提取和支持向量机分类结合起来，根据巴黎泌尿细胞学报告系统分为AI阴性、AI非典型和AI阳性三类。采用加权κ值和流行校正、偏倚校正κ值评估一致性。结果：在临床队列中，AI与组织病理学的符合率为83.0%，与专家细胞学的符合率为82.6%。加权κ值（κ, 0.631-0.661）反映了基本一致，差异主要在相邻类别之间。在健康筛查队列中，尽管患病率较低，但经患病率调整、偏倚调整的κ值为0.647。人工智能显示出99.7%的高阴性预测值和82.3%的阴性一致性，最小的假遗漏率（0.25%）。结论：人工智能辅助尿液细胞学显示了与专家解释和组织病理学标准的基本一致性。其筛除能力和高阴性预测值支持其作为可靠的分诊工具的潜力，在保持大规模筛查诊断可靠性的同时提高工作流程效率。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Cancer Cytopathology 医学-病理学

CiteScore

7.00

自引率

17.60%

发文量

130

审稿时长

1 months

期刊介绍： Cancer Cytopathology provides a unique forum for interaction and dissemination of original research and educational information relevant to the practice of cytopathology and its related oncologic disciplines. The journal strives to have a positive effect on cancer prevention, early detection, diagnosis, and cure by the publication of high-quality content. The mission of Cancer Cytopathology is to present and inform readers of new applications, technological advances, cutting-edge research, novel applications of molecular techniques, and relevant review articles related to cytopathology.